Norfloxacin removal from aqueous solution using biochar derived from luffa sponge

Pharmaceutical and personal care products (PPCPs) are a new type of widely produced trace pollutants that have become a global environmental problem and, accordingly, an active topic in dangerous material pollution and prevention research. Antibiotics are the most frequently detected PPCPs worldwide and the third most common drugs used in modern medicine, accounting for more than 6% of all prescription drugs (Boerner et al. 2003) after veterinary drugs, which account for more than 70% (Hallingsørensen et al. 1998). However, about 25–75% of animal-based drugs are excreted after being absorbed in metabolites and drug-targeted organs, such that the quantity and structure of antibiotic species in wastewater is rapidly becoming more numerous and complex. Such antibiotic contamination can increase microbial resistance, which can indirectly affect human health by causing a decline of the body's immune system capacity and increased allergic reactions (Kong et al. 2016).

Norfloxacin (NOR) is a ubiquitous antibiotic in the natural environment that is commonly used to treat enteritis dysentery because of its effectiveness against the DNA rotation enzymes (DNA gyrase) of pathogenic bacteria in the digestive tract (Holmes et al. 1985). Although environmental NOR concentrations are low, its chemical structure has specific biological implications with the potential to create cumulative adverse effects on non-target organisms. The removal of residue from the water is therefore necessary. Several methods to achieve NOR degradation (e.g., sonophotocatalytic degradation, ozonation, the solar Fenton method, photocatalytic degradation, and adsorption) have been investigated under a variety of experimental conditions (Alnajjar et al. 2007). Adsorption has numerous advantages over other conventional treatment methods because it is simple, inexpensive, and does not require the addition of nutrients (Lihme & Heegaard 2005).

Biochar (BC) is a solid product produced by the pyrolysis of biological organic material in an anoxic or anaerobic environment. BC can be used as a high-quality energy source, soil conditioner, reducing agent, slow-release carrier for fertilizers, and carbon dioxide sequestration agent. However, high costs are the main obstacle to using BC for large-scale wastewater treatment (Lehmann et al. 2011). The development of an affordable, broad-based, and rapidly renewable alternative adsorbent is therefore required (Spacie et al. 2011). Plant materials with unique environmental and cost advantages have been used to prepare new carbon materials with specific structures (Qi et al. 2016).

A wide range of plant material has been used to prepare BC (e.g., wild olive cores (Kaouah et al. 2013), papaya peel (Abbaszadeh et al. 2016), coffee grounds (Laksaci et al. 2017), and alligator weed (Kong et al. 2017)) and investigated under a variety of experimental conditions. The preparation of BC from luffa sponge as a raw material has been widely used in the adsorption of malachite green dye wastewater (Ang 2007), residual Cr(VI) (Miao et al. 2016), and phenol wastewater (Xiao et al. 2014); however, to the best of our knowledge, luffa sponge biochar for the removal of NOR residues in water has not been reported.

Luffa is widely cultivated in temperate and tropical regions around the world. Mature luffa forms a net fiber, called luffa sponge, that can be used for scrubbing and cleaning purposes as well as for medicinal use because of its diuretic properties to improve blood circulation and detoxification. Luffa is mainly composed of cellulose (82.4%), lignin (11.2%), and ash (0.4%), etc. (Tanobe et al. 2005). In addition, luffa has a uniquely porous physical structure, excellent mechanical strength, strong toughness, good acid and alkali properties (Akhtar et al. 2003; Vignoli et al. 2006; Ghali et al. 2009), and provides a good precondition for the preparation of biochar.

In this work, BC prepared from luffa sponge was used to adsorb NOR in synthetic wastewater. The BC was characterized based on its surface structure, specific surface area, porous structure, and Fourier-transform infrared (FTIR) spectrum (Mouille et al. 2003). Optimization of adsorption conditions was based on five factors: initial NOR concentration, pH, BC dosage, temperature, and contact time. Adsorption isotherm (Ozkaya 2006), adsorption kinetic (Chiron et al. 2003), and thermodynamic (Ghiorso & Sack 1995) models were used to study the adsorption mechanism.

Luffa sponge was purchased from a village in Jinan, China, and NOR (70458-96-7) from the Sangon Biological Engineering Corporation in Shanghai, China. The chemical formula of NOR is C₁₆H₁₈FN₃O₃, and its molecular weight is 319.33 g/mol. An initial NOR concentration of 100 mg/L was used.

Luffa sponge preparation first involved seed removal and cleaning with deionized (DI) water to remove dust. The luffa sponge was then dried at 70 °C for 2 days. Dried samples were crushed into 3- to 4-mm pieces using a high-speed grinder (HCP-100, Jinsui Company, Zhejiang). The broken samples were then infused with 85-wt% phosphoric acid at room temperature for 12 h at a solid-liquid ratio of 1:4 (g:mL) (Girgis & Ishak 1999). After activation, the samples were carbonized for 2 h in a box-type resistance furnace at 450 °C (Beijing Yongguang Company). The sample was then washed with DI water until the pH was close to 7. Once cooled, the sample was dried at 70 °C for 48 h and ground to a 200 mesh.

The surface area structure of the BC was observed using scanning electron microscopy (SEM) (Ory et al. 2006). The BC surface area and pore size were analyzed by Brunauer–Emmett–Teller (BET). The effect of adsorption on BC functional groups was analyzed by FTIR. The BC sample and KBr powder were mixed evenly in the desired proportions, and then pressed into a piece, and scanned over the 4,000–400 cm⁻¹ wavelength range.

Different concentrations of NOR (50–200 mg/L) were mixed with a fixed amount of BC in a 100-ml conical bottle. The sample was stirred at a constant stirring speed of 180 rpm in an oscillator at a controlled temperature of 288, 298, or 303 K. The pH was regulated during the experiments using titrated Na(OH) and HCl to pH = 6.

Figures 1(a) and 1(b) were used to characterize the surface properties of luffa BC by SEM. Figure 1(a) clearly shows that the BC surface is rough and has a rich pore structure that forms after the activation of phosphoric acid when luffa undergoes high-temperature carbonization, which leads to internal etching and enhances bio-carbon adsorption. However, the SEM adsorption spectrum (Figure 1(b)) shows a reduced pore structure because the BC surface was filled by NOR molecules, which indicates that luffa BC is a good NOR adsorbent.

The BC pore structure distribution (Figure 2(a)) shows that most pores are mesopores (diameter: 2–10 nm), in addition to micropores (diameter: 0–2 nm). However, no macropores (diameter: 10–50 nm) were observed on the BC surface. The BET surface area of the BC was 822.35 m²/g, and the mean aperture was 5.35 nm. The specific surface area and pore size have a strong and generally inverse relationship (Graber et al. 2012). The N₂ adsorption/desorption isotherm of the BC (Figure 2(b)) exhibits a curve representative of a mixture of BC forms I and IV. This indicates that the gas adsorption rate rises quickly to a limit with increasing pressure in the low-pressure range, and a wide hysteresis loop appears at high pressure (Wang et al. 2013). This also indicates that the porous BC structure is a mixture of micropores and mesopores, which is advantageous for adsorption.

Figure 3(a) shows the effect on adsorption capacity and removal rates during a 0–360-minute contact time. The removal rate rises rapidly from 59% (±0.97) to 83% (±0.37) in the first 60 minutes because of the adsorption of NOR molecules onto vacant active sites, which also provide a high driving force for the adsorption process in general (Kong et al. 2017). After 240 minutes, the removal rate becomes constant, indicating that adsorption has reached a balanced state. This occurs because the number of empty active sites gradually decreases with time, interaction between NOR molecules and internal particles produces repulsion (Nagda & Ghole 2009), and the BC surface becomes saturated with adsorbed NOR molecules (Li et al. 2009). The equilibrium time of BC adsorption of NOR is therefore 240 minutes.

Figure 3(b) shows the effect of temperature on the NOR removal rate and adsorption capacity. The removal rate and adsorption amount decrease slightly with increasing temperature from 96% (±0.80) to 95% (±0.60) and from 135 (±0.45) to 131 (±0.36) mg/g, respectively, which indicates that the adsorption process is exothermic (Kuzmichov & Pogorelov 2003). Previous studies showed that the adsorption properties of adsorbent molecules are enhanced under high-temperature conditions, which also supports that adsorption is exothermic. Adsorption efficiency tests can therefore be performed at room temperature.

The effect of BC dosage on the NOR removal rate and equilibrium adsorption amount is shown in Figure 3(c). Increasing the BC dosage from 0.1 to 0.6 g/L results in a rapid initial increase of the removal rate from 35% (±0.30) to 95% (±0.45). This implies that the number of adsorbed sites increases with increased amounts of BC, which reduces competition between NOR molecules (Xu et al. 2017). However, equilibrium adsorption decreases with increasing removal rate because as the removal rate approaches 100%, the limited remaining NOR solute molecules cannot be adsorbed onto more active sites, which results in more active-site vacancies (Kong et al. 2017). An adsorbent concentration of 0.5 g/L was therefore selected to maximize the effectiveness of luffa sponge.

The initial aqueous solution pH places important controls over the adsorption process because it influences the surface charge and BC morphology (Cheng et al. 2008). The results indicate an optimum pH range of 5–7 (Figure 3(d)), for which the maximum adsorption capacity and removal rate were 128 mg/g (±0.300703) and 91% (±0.520833), respectively. The point of zero charge (pH_pzc) is an important parameter that indicates the influence of pH on adsorbent chemical characteristics and adsorption capacity. The determination of PZC of BC is shown in Figure 3(e). The pH_pzc for the BC was found to be about 1.8, Thereafter, the dissociation constant for the fluoroquinolones antibiotics have two ionization constant values at 5.45 and 6.2 (Bhatia et al. 2016). The cationic type dominates when pH <5.45 (László & Szűcs 2001), while the anionic type is dominant in pH >6.2. While the maximum adsorption capacity of NOR was at pH 6, the predominant species is the zwitterionic form. The adsorption of NOR was through electrostatic interactions and cation exchange, respectively (Adriano et al. 2006). This explains the improved adsorption at pH = 6.

Figure 3(f) shows that the removal efficiency decreases with increasing initial concentration from 92% (±0.601407) to 77% (±0.68306), and the adsorption capacity increases from 39 (±0.521301) to 121 (±0.491926) mg/g. The removal efficiency decreases because of the limited number of empty active sites and saturation at low concentrations. A large number of molecules therefore compete for limited surface sites (Qi et al. 2016). The adsorption quantity increases with increasing initial NOR concentration, but the adsorption efficiency does not improve because the number of active sites is insufficient to meet the demand of such a large number of molecules. An initial concentration of 30 mg/L was therefore used in subsequent experiments.

The adsorption kinetics data were fitted with pseudo-first-order, pseudo-second-order, and intraparticle diffusion models to determine the adsorption mechanism. The optimal adsorption kinetic model was determined by a combination of the normalized standard deviation Δq (%) and correlation coefficient R². Fitted parameters of the three kinetic models are listed in Table 3. The correlation coefficient (R²> 0.99) simulated by the pseudo-second-order model is higher than those of other models, and the Δq values are smaller than those of other models. Additionally, the results show that the pseudo-second-order model (Q_e) best fits the Q_exp data, indicating that the pseudo-second-order model can well fit the adsorption process of the BC to NOR solution. Thus, the adsorption rate is controlled by chemical adsorption of valence forces by exchanging or sharing electrons between the adsorbent and NOR molecules (Rogachev 2008; Mahmoud et al. 2016).

Calculated Gibbs free energy, entropy, and enthalpy values are listed in Table 4. The ΔG indicates that the degree of spontaneity and adsorption of NOR were feasible, and more negative ΔG values produced a better adsorption effect, indicating spontaneous adsorption (Abbaszadeh et al. 2016). The free energy decreases with increasing temperature, which leads to decreased adsorption rates at high temperature. The negative ΔH value confirms that adsorption is exothermic, which is consistent with the increased adsorption capacity observed in the Langmuir isotherm model with decreasing temperature. Additionally, the temperature effect also confirmed that the BC adsorption NOR process is exothermic process. The ΔS value is negative, which also indicates that randomness reduces adsorption.

The FTIR spectra of BC before and after NOR adsorption are shown in Figure 4. The types of surface functional groups changed very little before and after BC adsorption. The spectral peak at 400–500 cm⁻¹ was observed before adsorption, due to the vibration of a metal-oxygen and metal-hydroxyl, but disappeared after the NOR biosorption by BC (Puziy et al. 2003). The peak at 1,080–1,800 cm^–1 indicates that the surface of the bio-carbon contains phosphorus functional groups due to pretreatment with phosphoric acid (Kaouah et al. 2013). The absorption peak at 1,570 cm⁻¹ is caused by contraction vibrations of the aromatic ring. The 1,600 cm⁻¹ peak is related to tensile vibration of the benzene ring or C = C bond (Shanmugharaj et al. 2007). The peak at 1,700 cm⁻¹ is due to the vibration of C = O bonds in the carboxyl or conjugate carbonyl group (Seredych et al. 2009). Absorption peaks at 3,420 cm⁻¹ are derived from the elastic vibrations of –OH bonds in the carboxyl, phenol hydroxyl, and water molecules (Laksaci et al. 2017). Figure 4 shows that the peak strength of BC decreased slightly after adsorption. It is shown that the surface groups of BC are involved in the process of adsorption of NOR, and they interact with NOR or are masked by NOR molecules.

Luffa was widely cultivated in developing countries as an environmentally friendly biomaterial in eastern Asia and South America (Demir et al. 2008). Because of the high yield and easy availability of luffa, luffa sponge can be prepared into low cost biochar. Commercial activated carbon is usually used to remove contaminants from water. However, commercial activated carbon is expensive, leading to a higher cost of wastewater treatment (the cost of commercial activated carbon is CN¥ 10.4/kg). The study of our research team showed the cost of production of luffa sponges biochar is CN¥ 2.1/kg. Therefore, use of luffa biochar can greatly reduce the cost of the water treatment field. This greatly reduces the cost of water treatment areas and facilitates economic diversification in developing countries. In addition, there are many areas where biochar can be applied. Biochar can also improve soil fertility (Yuan et al. 2018) and soil mechanisms (Khan et al. 2016) while treating soil contaminants (Khorram et al. 2018). Moreover, biochar can affect the mitigation of greenhouse gas emissions (Li et al. 2018). Thus, BC has great significance and wider application prospects in global carbon biogeochemical cycling, mitigation of global climate change research and ecological remediation of pollutants.

BC derived from luffa sponge is an effective low-cost material for removing NOR from aqueous wastewater. SEM analyses show that the adsorbent has a coarse surface with a well-developed porous structure containing abundant micropores and mesopores. The BC surface area is high (822.35 m²/g), and the average pore size is 5.35 nm. These physical characteristics indicate that BC has excellent properties for NOR adsorption. The maximum adsorption capacity of the BC was 250 mg/g. Batch adsorption studies were performed to evaluate the ability of BC to remove NOR from aqueous solutions. The NOR adsorption kinetic data correlate well with a pseudo-second-order kinetic model, with a high R² (>0.99), which shows that chemisorption is the rate-limiting factor. The experimental data are best described by the Freundlich isotherm model (R² > 0.99), which indicates multilayer adsorption. The thermodynamic parameters show that adsorption is exothermal and spontaneous. The FTIR spectra indicate that the BC surface has more acidic oxygen-containing groups such as carboxyl, phenol hydroxyl, lactone, and carbonyl. Because luffa sponge is inexpensive and easy to obtain, this technology offers broad application prospects to manage wastewater clean-up.

This work was supported by the International Postdoctoral Exchange Fellowship Program (No. 20180063), National Natural Science Foundation of China (No. 51708340), Major Science and Technology Program for Water Pollution Control and Treatment (2017ZX07101-001), Promotional Research Fund for Excellent Young and Middle-aged Scientists of Shandong Province (ZR2016CB18 and ZR2015DM012) and Project of Shandong Province Higher Educational Science and Technology Program (No. J15LE07 and No. J14LD02).